Pyro
Build & refine probabilistic ML models with deep neural networks via open-source Pyro platform, backed by Broad Institute & Linux Foundation.
Pyro is an open-source deep universal probabilistic programming platform developed by a community including Broad Institute and originally from Uber AI, featuring a flexible language for defining generative models, automatic differentiation, and support for TensorFlow and PyTorch. Notable achievements include time-to-event modeling, GANs, and advanced machine learning tasks. Pyro offers tools and services to build probabilistic ML models with deep neural networks using AI for statistical inference on complex data distributions, including a language, automatic differentiation, TensorFlow/PyTorch support, and extensive resources.
About Pyro
Pyro is a platform for deep universal probabilistic programming developed and maintained by a community of contributors, including a team at the Broad Institute. Originally developed at Uber AI, Pyro became a project of the Linux Foundation in 2019.
Pyro provides tools and services to build probabilistic machine learning models using deep neural networks. It leverages artificial intelligence (AI) to perform statistical inference on complex data distributions, allowing for more accurate and robust model development. Pyro's core offerings include:
- A flexible and expressive probabilistic programming language for defining generative models.
- Automatic differentiation for efficient computation of gradient-based optimization algorithms.
- Support for various neural network architectures and deep learning libraries, such as TensorFlow and PyTorch.
- Extensive documentation, tutorials, and examples to help users get started with the platform.
Pyro has achieved notable achievements in the deep learning community. It is used in various applications, including time-to-event modeling, generative adversarial networks (GANs), and other advanced machine learning tasks. Pyro's open-source nature enables a vibrant community of contributors and users to continue advancing the platform's capabilities.